Second Circuit affirms rejection of "All Natural" survey as too leading

Bustamante v. KIND, LLC, 100 F.4th 419 (2d Cir. 2024)

The court of appeals affirmed summary judgment in favor of KIND on Bustamante’s false advertising consumer protection class action claims based on KIND’s “All Natural” labeling. The complaint alleged that eleven ingredients contained in some relevant KIND products were “non-natural”: Soy Lecithin; Soy Protein Isolate; Citrus Pectin; Glucose Syrup/“Non GMO” Glucose; Vegetable Glycerine; Palm Kernel Oil; Canola Oil; Ascorbic Acid; Vitamin A Acetate; D-Alpha Tocopheryl Acetate/Vitamin E; and Annatto.

Eventually, the district court excluded plaintiffs’ survey and scientific experts and granted summary judgment.

“To establish deception under the reasonable consumer standard at the summary judgment stage, plaintiffs must present admissible evidence establishing how the challenged statement – ‘All Natural’ – tends to mislead reasonable consumers acting reasonably.”

Although errors in survey methodology generally go only to weight rather than admissibility, it was not an abuse of discretion to exclude the survey expert here. The court found that the survey “does not assist the trier of fact because it is biased, leading, and to the extent it provides any insight, cannot provide the objective standard necessary to answer the key question in this case.”

The survey surveyed California, Florida, and New York consumers who had purchased KIND products, or products from a KIND competitor, in the last twelve months. Respondents saw “a mock-up of the front of a brand-neutral product package and [were instructed] to ‘examine it like you were shopping’ ” and “to assume that the nutrition snack bar is a ‘popular national brand.’ ” The mock-up label displayed the words “All Natural,” and in several respects resembled the packaging of a KIND bar.

The first relevant question asked: “Because of this descriptor [All Natural], what is your expectation for this product?” It offered three possible choices: (a) “Will NOT contain artificial and synthetic ingredients;” (b) “Will contain artificial and synthetic ingredients;” or (c) “Not sure/No expectation.” “86.4% of consumers expected the Product with the ‘All Natural’ claim ‘will NOT contain artificial and synthetic ingredients.’ ” The survey did not define the terms “artificial” or “synthetic.”

The district court found that this question didn’t help determine “in any meaningful sense how reasonable consumers understand the ‘All Natural’ claim, or to test plaintiffs’ theory.” It was “biased” and “lead[ing]” because it “improperly directs survey participants to the ‘correct’ answer” and “is plainly designed to validate plaintiffs’ theory” of liability. This characterization was not manifestly erroneous, especially because the expert conceded that he “worded [his] substantive response options on the basis of [his] understanding of the Plaintiffs’ theory of liability.” The Second Circuit has previously held that a plaintiff could not rely on a survey based on a question that, like this one, “was an obvious leading question in that it suggested its own answer.” (Citing Universal City Studios, Inc. v. Nintendo Co., 746 F.2d 112 (2d Cir. 1984), where the question was “To the best of your knowledge, was the Donkey Kong game made with the approval or under the authority of the people who produce the King Kong movies?” That’s a far more leading question; the sin here seems to have been that the question was closed-ended, even though it had a don’t know/not sure option. Would the Second Circuit be ok with starting with an open-ended question? With asking people whether “All Natural” means “no artificial/synthetic ingredients”? What else could you possibly ask to test plaintiffs’ theory of liability?)

The choice “to display the ‘All Natural’ claim in isolation, rather than as part of the ‘All Natural/Non GMO’ statement, as it always appeared on KIND labels” further undercut the relevance of the results.

The second question asked: “Because of this descriptor [All Natural], what is your expectation for this product?” The options were: (a) “Is NOT made using these chemicals: Phosphoric Acid, Hexane, Potassium Hydroxide, Ascorbic Acid”; (b) “Is made using these chemicals: Phosphoric Acid, Hexane, Potassium Hydroxide, Ascorbic Acid”; or (c) “Not sure/No expectation.” The survey didn’t describe or define these “chemicals” (court’s scare quotes). The results were similar: over 76% of respondents chose (a)

It was also not manifestly erroneous to find this irrelevant. By providing a list, the survey “led survey participants down the path of selecting the answer preferred by plaintiffs.” (Would it have been better to give them an actual ingredient list?) Also, by listing the “chemicals” without defining them, the survey failed to differentiate between “ascorbic acid,” a form of Vitamin C safe for human consumption, and “phosphoric acid,” which is “not safe for ingestion.”

Likewise, it was not an abuse of discretion to exclude the scientific expert because he wouldn’t assist in identifying what reasonable consumers considered artificial or synthetic. He developed a framework that “examined each ingredient’s origin, the extent to which the ingredient had been processed from its natural form, and the final form of the ingredient.” He opined on whether the ingredients could be classified as “natural” under his framework, but didn’t apply a definition used elsewhere, including in the complaint or by the survey. Nor did he specifically analyze KIND ingredients, only how they were “typically” sourced. “But, without some evidence to the contrary, there is no reason to assume that [the expert’s] personal understanding of the term ‘natural’ is relevant to how a reasonable consumer would understand that same term.” Because of that flaw, “the report adds no useful information that would help the trier of fact determine the answer to the relevant legal question: whether consumers were actually deceived.”

Without the expert evidence, summary judgment for KIND was appropriate. Named plaintiffs’ own testimony wasn’t enough because they didn’t provide a cohesive definition of what “All Natural” meant, whether it would mean “containing no artificial or synthetic ingredients, or what it means to be artificial or synthetic. Plaintiffs’ depositions instead showed how variable definitions of “All Natural” can be:

For example, one plaintiff testified that she expected “All Natural” to mean not synthetic. Another plaintiff testified that she expected “All Natural” to mean that the product was made from whole grains, nuts, and fruit. Yet another explained her belief that “All Natural” meant that the ingredients were literally plucked from the ground. Notably, several plaintiffs testified that consumers could have different understandings about the implications of the term “All Natural,” that these understandings could change over time, and that not everyone would agree with their particular understanding of that term. Plaintiffs fail to explain how a trier of fact could apply these shifting definitions to reach a conclusion as to whether the use of the term “All Natural” on KIND product labels was deceptive.

KIND’s own internal documents weren’t helpful because all they showed was that KIND had its own conception of the term, but didn’t show what a reasonable consumer’s understanding was. (Courts used to be more willing to say “the seller’s beliefs about what its audience wants are good circumstantial evidence, given the seller’s incentives,” and they still do in trademark cases.) The FDA’s own request for comments also demonstrated lots of varied understandings.

Nor was it enough for plaintiffs to use the dictionary. (That’s just for courts.) The definition identified, “existing in or caused by nature; not made or caused by humankind,” “is not useful when applied to a mass-produced snack bar wrapped in plastic. Such a bar is clearly made by humans.”